SAIMPLE — AI Model Validation Platform
numalis
Static analyzer for AI model validation, explainability, and robustness — compliant with ISO/IEC 24029-2 and EU AI Act Articles 13 & 15. Available as SaaS or on-premises for critical-sector aerospace and defense applications.
Technical specifications
- Standards compliance
- ISO/IEC 24029-2, EU AI Act Articles 13 & 15
- Validation method
- Abstract interpretation (formal methods)
- Model types
- Neural networks, SVMs
- Deployment modes
- SaaS (web + API), On-premises, Educational
- Target sectors
- Aerospace, defense, nuclear, rail, banking
- Applications
- Image processing, predictive maintenance, data augmentation robustness
About
SAIMPLE (Static Analyzer for AI Validation) is Numalis’s core AI validation platform built on abstract interpretation formal methods. It enables data scientists and AI engineers to validate neural network robustness and generate explainability reports in compliance with ISO/IEC 24029-2 and EU AI Act Articles 13 and 15. The tool supports the full AI development lifecycle from dataset quality analysis to pre-production reliability validation, providing risk-reward trade-off analysis for neural networks and SVMs and automating compliance documentation for critical-sector applications. SAIMPLE is particularly relevant to aerospace, defense, nuclear, and rail sectors where safety-critical AI must be formally verified before deployment. Deployment options include a web-based SaaS with API access and a secure on-premises installation for regulated industries.
Documentation
No public datasheet yet — request the datasheet / ICD from the supplier.